Research of classification and prediction model in real-time representation of neurons

To research the intrinsic pattern of the neuron firing and deal with the classical linear discriminant analysis could not effectively classify the neuron sampling data.The regularization linear discriminant analysis and nearest shrunken centroids are put forward.And a new analysis method of neuron frequency is designed according to the characteristic of the neurons.Ultimately,the efficiency of the proposed methods is verified through the cross validation.Experiments shows that the validity of the proposed methods,prove that the intrinsic pattern of the neuron firing and the feasibility of the classification and prediction to the different stimuli through the analysis of the ensemble firing patterns.